Hi all!
Without very very precise and distortion free measurement equipment difficult to find out. Maybe someone here have some kind of spectrum analyzer to check this out. I have this concern with my DSP active crossover as of the miniDSP 2 X 4 HD with listening tests with a high SPL super tweeter. Depending of the crossover frequency the time domain and frequency domain signal portion of the super tweeter channel get very low in amplitude. With 7 to 9 kHz crossover with lower listening level i sense putting my ear very close to the horn super tweeter with about 100db / meter sensitivity only a kind of contact noise and nothing else that reminds me of that what you can hear when you have massive quantization distortion. Have anybody figured it out with serious measurement equipment? Just apply a 443 Hz standard pitch square wave signal with smaller signal level in steps - let us say with 40, 30, 20,10,0 dB attenuation of the maximum input level and measure the small signal distortion level on the analog output of the DSP crossover with a crossover point in the steps let's say 6 kHz , 7 kHz, 8 kHz, 9 kHz and 10 kHz. For the low pass output no problem, for the high pass output the signal portion in my theory is so small that the A/D - PEQ - XO - volume attenuation - D/A section of the DSP crossover introduce a relatively significant amount of distortion to the high pass filtered small signal portion of the square wave signal. There are two variants of that with level attenuation already of the input signal in front of the A/D section and DSP based volume attenuation for the output level in front of the D/A section, this will be a significant difference for higher attenuation than for low attenuation - 0 db attenuation no difference, 40 dB attenuation bigger difference. I am a non native English speaker and i hope that i get the things explained in more or less plain English. Feedback is very appreciated, many thanks in advance.
P.S.: having said that i think the discussion whether a FIR or a IIR filter reconstruct better the original shape of a square wave signal is irrelevant for normal listening levels of the music program 30 or 40 dB below the maximum output of the system unless you have the analog volume attenuation between crossover and amp - definitely NOT what you have in a plate amp like the Hypex Fusion
Without very very precise and distortion free measurement equipment difficult to find out. Maybe someone here have some kind of spectrum analyzer to check this out. I have this concern with my DSP active crossover as of the miniDSP 2 X 4 HD with listening tests with a high SPL super tweeter. Depending of the crossover frequency the time domain and frequency domain signal portion of the super tweeter channel get very low in amplitude. With 7 to 9 kHz crossover with lower listening level i sense putting my ear very close to the horn super tweeter with about 100db / meter sensitivity only a kind of contact noise and nothing else that reminds me of that what you can hear when you have massive quantization distortion. Have anybody figured it out with serious measurement equipment? Just apply a 443 Hz standard pitch square wave signal with smaller signal level in steps - let us say with 40, 30, 20,10,0 dB attenuation of the maximum input level and measure the small signal distortion level on the analog output of the DSP crossover with a crossover point in the steps let's say 6 kHz , 7 kHz, 8 kHz, 9 kHz and 10 kHz. For the low pass output no problem, for the high pass output the signal portion in my theory is so small that the A/D - PEQ - XO - volume attenuation - D/A section of the DSP crossover introduce a relatively significant amount of distortion to the high pass filtered small signal portion of the square wave signal. There are two variants of that with level attenuation already of the input signal in front of the A/D section and DSP based volume attenuation for the output level in front of the D/A section, this will be a significant difference for higher attenuation than for low attenuation - 0 db attenuation no difference, 40 dB attenuation bigger difference. I am a non native English speaker and i hope that i get the things explained in more or less plain English. Feedback is very appreciated, many thanks in advance.
P.S.: having said that i think the discussion whether a FIR or a IIR filter reconstruct better the original shape of a square wave signal is irrelevant for normal listening levels of the music program 30 or 40 dB below the maximum output of the system unless you have the analog volume attenuation between crossover and amp - definitely NOT what you have in a plate amp like the Hypex Fusion
Last edited by a moderator:
Did you say the power amplifier is class-D? If so, it might be culprit more than the miniDSP, although both could have some possible effects.
https://purifi-audio.com/2020/04/28...'t something,but not entirely random moments.
https://purifi-audio.com/2020/04/28...'t something,but not entirely random moments.
I could minimize the issue after putting the volume level attenuation between analog output of the DSP crossover and the analog input of the power amp with a second hand home theater preamplifier where i use the 7.1 channel input and the analog volume attenuator. I had to tweak around to get the crossover analog input only a few dB under the maximal allowed input of the miniDSP and bypass the DSP based volume attenuation of the crossover itself.
Amplifier is the power section of an Marantz AVR for the moment - no class D as far as i know. I will purchase two multi channel power amplifier as soon as the room for the home theatre is ready in our new home and the AVR is needed there again.
Amplifier is the power section of an Marantz AVR for the moment - no class D as far as i know. I will purchase two multi channel power amplifier as soon as the room for the home theatre is ready in our new home and the AVR is needed there again.
Last edited by a moderator:
So maybe the miniDSP outputs too much voltage for the power amplifier input? Or, maybe the attenuator is acting as LP filter of quantization and or other audible noise/distortion? Something else?
The issue was clearly perceptible with lower listening levels putting the ear close to the super tweeter and 8 - 10 kHz crossover frequency and the volume level attenuation in front of the DSP crossover already done on the analog output of the audio network streamer that i have for the multi room stuff of my ROON music server. I have also some kind of reference acoustic output with the Genelec satellites and subwoofer stuff at the desk of my home office with a very clear sound from lowest to highest music program listening levels.
You are going into the analog input of the miniDSP? If so that's not very desirable from a noise and distortion perspective. Better to use a digital input.
I would like to, but i use a double mono 1 X 4 R / L setup with two miniDSP 2 X 4 HD - i do not have the foggiest idea how to split the digital signal in two separated ones for the right channel miniDSP and one for the left channel miniDSP
Anything help at: https://www.google.com/search?q=how...0j33i299l2.12436j1j1&sourceid=chrome&ie=UTF-8
At least it might be nice if you could try it and see what you think.
At least it might be nice if you could try it and see what you think.